Objective: To investigate the factors which affect the motor evoked potential (MEP) responsiveness and parameters and to find the correlation between the function of the upper extremities and the combined study of MEP with a diffusion tensor tractography (DTT) in patients with stroke. Methods: A retrospective study design was used by analyzing medical records and neuroimaging data of 70 stroke patients who underwent a MEP test between June 2011 and March 2013. MEP parameters which were recorded from the abductor pollicis brevis muscle were the resting motor threshold, latency, amplitude, and their ratios. Functional variables, Brunnstrom stage of hand, upper extremity subscore of Fugl-Meyer assessment, Manual Function Test, and the Korean version of Modified Barthel Index (K-MBI) were collected together with the biographical and neurological data. The DTT parameters were fiber number, fractional anisotropy value and their ratios of affected corticospinal tract. The data were compared between two groups, built up according to the presence (MEP-P) or absence (MEP-N) of MEP on the affected hand. Results: Functional and DTT variables were significantly different between MEP-P and MEP-N groups (p<0.001). Among the MEP-P group, the amplitude ratio (unaffected/affected) was significantly correlated with the Brunnstrom stage of hand (r=-0.427, p=0.013), K-MBI (r=-0.380, p=0.029) and the time post-onset (r=-0.401, p=0.021). The functional scores were significantly better when both MEP response and DTT were present and decreased if one or both of the two studies were absent. Conclusion: This study indicates MEP responsiveness and amplitude ratio are significantly associated with the upper extremity function and the activities of daily living performance, and the combined study of MEP and DTT provides useful information.